Short answer: focus a WebStorm text field (AI Assistant chat/prompt box, editor comment block, commit message field, or code review note), press ⌃+⌥+R (Mac) or Ctrl+Alt+R (Windows/Linux), speak, and press again to insert.
Why WebStorm Gains From Voice
WebStorm removes friction from code edits, refactors, navigation, and inspections. The bottleneck is usually explanatory prose: AI prompts with constraints, TSDoc comments that encode intent, and commit bodies that explain trade-offs.
Voice is useful because those fields are numerous and contextual. You can dictate while the relevant diff, file, or inspection is visible.
WebStorm-Native Flows With Real Depth
AI Assistant prompts with file-scoped constraints
In the AI Assistant panel, dictate one structured prompt instead of iterative one-liners:
"Refactor src/services/tokenService.ts only. Keep public API unchanged. Replace duplicate retry logic with shared helper. Add unit tests for timeout and invalid-token branches. Do not touch UI files."
This gives the assistant scope, non-goals, and test expectations in one pass.
TSDoc/JSDoc for exported APIs
WebStorm can auto-generate doc stubs, but high-quality docs still require intent and edge cases. Click inside a comment block and dictate:
"Parses webhook payload and normalizes provider-specific event names. Throws WebhookSignatureError on invalid signature. Idempotent for duplicate delivery IDs."
Then add @param, @returns, and @throws tags quickly.
Commit tool window with rationale while diff is visible
WebStorm's commit UI is strongest when message + diff are reviewed together. Dictate both subject and body while scanning staged changes:
"Extract auth retry policy into shared helper to avoid divergence between REST and GraphQL clients. Adds regression tests for 429 backoff. No behavior change for successful requests."
This improves review quality compared with generic one-line commits.
TODO/FIXME conversion into actionable notes
When you hit unclear logic, dictate a concrete TODO in place:
"TODO: Replace hardcoded currency map with server-provided metadata before multi-region rollout. Current map breaks JPY rounding."
These comments are much more searchable and actionable.
HTTP client request docs in .http files
WebStorm teams often store API requests in .http files. Dictate request notes above examples:
"Use sandbox token only. Endpoint rate-limited to 20 RPM. Response schema differs in production for discounts array."
This saves future debugging time.
AI Inputs In WebStorm
Prompt boxes where AICHE helps most:
- AI Assistant chat/composer prompts
- follow-up refine prompts after initial output
- inline explain/refactor prompts from selected code
AICHE inserts prompt text. WebStorm AI tools execute generation and edits.
How It Works
- Open the exact WebStorm text field.
- Press ⌃+⌥+R / Ctrl+Alt+R.
- Speak full context with scope and constraints.
- Press again to insert.
- Review and run your normal inspections/tests.
Audio is streamed for cloud transcription, processed, and discarded immediately after processing, within 1 second. No persistent audio copy.
FAQ
Need a WebStorm plugin?
No.
Can it handle identifiers and project jargon?
Yes. Use Software Development profile plus Custom Vocabulary for project-specific terms.
Wayland?
Usually yes; some desktop environments need permission configuration.
Try it now: open AI Assistant in WebStorm and dictate one scoped refactor prompt with explicit file boundaries, non-goals, and test expectations.